Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Interjudge Agreement, Self-Enhancement, and Liking: Cross-Cultural Divergences Steven J. Heine University of British Columbia Kristen Renshaw University of Pennsylvania Please address correspondence to: Steven J. Heine 2136 West Mall, University of British Columbia Vancouver, BC, V6T 1Z4 Canada Tel: (604) 822-6908. Fax (604) 822-6923 E-mail: [email protected] DRAFT COPY: PLEASE DO NOT QUOTE WITHOUT PERMISSION Self-peer evaluations across cultures 2 Abstract We investigated whether the lower self-enhancement found among Japanese is due to them being more accurate in their self-perceptions than Americans. Japanese and American participants were recruited from school clubs, where groups of 5 people rated each other and themselves. Replicating some past research, the Japanese sample was overall self-critical (i.e., people rated themselves less positively than they were rated by their friends) whereas the American sample was overall self-enhancing. Moreover, as the desirability of the traits increased, Americans showed more self-enhancement whereas Japanese showed more self-criticism. An accuracy account is unable to account for the cultural differences in self-enhancement as Americans showed more accuracy in their self-perceptions (as evidenced by self-peer agreement) than Japanese. Intracultural analyses further revealed that individual self-enhancement can be “unpackaged” by trait measures of independence and interdependence. Exploratory analyses of liking were also conducted, revealing that American liking hinged on perceived similarity, self-verification, familiarity, and reflected-self-enhancement, whereas Japanese liking was based on familiarity, reflected self-enhancement, lower independence, and interdependence. Self-peer evaluations across cultures 3 Much recent research investigating cultural influences on the self-concept has focused on cultural comparisons of self-enhancing tendencies between East Asians (particularly Japanese) and North Americans. Research conducted with North Americans finds tendencies for individuals to view themselves in unrealistically positive terms (e.g., Dunning & Cohen, 1992; Weinstein, 1980), in contrast, findings with Japanese are much weaker or absent. For example, research on the better-than-average effect (Markus & Kitayama, 1991), unrealistic optimism (Heine & Lehman, 1995), and self-serving attributional biases (Kitayama, Takagi, & Matsumoto, 1995) consistently reveals that Japanese demonstrate far less evidence of these biases than North Americans. The convergent evidence demonstrates that Japanese are less inclined to self-enhance than North Americans (Heine, in press-a; Heine, Lehman, Markus, & Kitayama, 1999). It is important to consider why these cultural differences exist. One perspective is that self-criticism is more functional for Japanese, whereas self-enhancement is more functional for North Americans, because of the different lay theories of the self that are commonly embraced in the two cultures (e.g., Heine, in press-a; Heine et al., in press; Kitayama, Markus, Matsumoto, & Norasakkunkit, 1997). Self theories that are more commonly shared among Japanese than among North Americans include the beliefs that individual selves are malleable and incremental (e.g., Heine, in press-b; Heine et al., in press). These views of self are associated with an understanding of the self as adjustable (e.g., Morling, Kitayama, & Miyamoto, in press) and ultimately improvable through efforts (Heine et al., in press; Holloway, 1988). This reasoning suggests that self-critical views are more functional for Japanese compared with North Americans as they highlight areas upon which individuals can improve. An alternative explanation deserves consideration. That is, the weaker self-enhancing motivations among Japanese might be due to them being better attuned to their actual standings. Self-peer evaluations across cultures 4 There is some evidence consistent with the notion that Japanese self-assessments should harmonize with how they are evaluated by others. For example, much psychological and ethnographic research describes East Asians as possessing a heightened sensitivity to how an audience is viewing them, and of maintaining face (Chang & Holt, 1994; Cohen & Gunz, 2001; Lebra, 1976). This heightened sensitivity to how the self appears to others suggests that Japanese should be more likely than North Americans to derive their self-evaluations from how they feel they are being viewed by others and thus evince higher levels of interpersonal accuracy. That is, cultural differences in self-enhancement might not be due to self-critical motivations on the part of Japanese, but to their greater concerns with interpersonal accuracy. Past cross-cultural research on self-enhancement has been unable to test for cultural differences in interpersonal accuracy as the methodologies that have been employed are only able to identify the existence of self-enhancement at the group level. That is, people as a group might rate themselves in unrealistically positive ways (e.g., on average individuals might estimate that they are better than the average other), but this does not tell us which individuals are self-enhancing (Colvin, Block, & Funder, 1995; John & Robins, 1994). Some individuals really are better than average, upwards of 50% in fact, and these group measures confound self-enhancement and competence at the individual level. Without an objective benchmark with which to compare the individual we cannot determine how accurate individuals’ self-assessments are (Colvin et al., 1995), and thus cannot explore whether accuracy varies cross-culturally. We sought to investigate whether Japanese differ in the accuracy of their self-perceptions compared with Americans. Determining accuracy requires measuring how much evaluations depart from a standard. The standard that we adopted in the present study was one used in prior research by John and Robins (1993): namely, how individuals were evaluated by four individuals who knew them very well. These four individuals were fellow members of the participants’ Self-peer evaluations across cultures 5 groups at university. The extent to which self-assessments were similar to the average assessment by four of their peers provides an index of accuracy. The Bases of Liking across Cultures Although the primary purpose of this study was to investigate self-enhancement and accuracy, the study’s targeted samples of close-knit groups yielded a unique set of data that also allowed for an exploration of the bases of liking across cultures. In addition to the various self and peer evaluations that participants made of their 4 other group members, participants were also asked to indicate how much they liked each of their peers, providing a set of data that accommodated empirical tests of various theories of interpersonal attraction. To date, aside from research on romantic love (e.g. Dion & Dion, 1993), there is scant research on cultural differences in the basis of liking. Humans are a social species, and thus it would seem that interpersonal attraction is a universal concern, however, culturally divergent construals of self may influence the bases on which this attraction is founded. Much research conducted in North America has identified a few bases of liking: 1) the similarity effect (e.g., Byrne, 1971), which states that we like those who are similar to us, largely because we find being with those who are like us rewarding; 2) self-verification, which states that we come to prefer people who view us similarly to how we view ourselves so that a consistent self can be maintained (e.g., Swann, Stein-Seroussi, & Giesler, 1992); 3) the mere exposure effect (e.g., Zajonc, 1968), which states that we come to prefer people who are familiar to us through frequent interaction because familiarity reduces response competition and facilitates ease of recall (e.g., Harrison, 1977); and 4) what we term “reflected self-enhancement,” the notion that we prefer people who view us in positive terms as a means to feel better about ourselves (e.g., Murray, Holmes, & Griffin, 1996). Self-peer evaluations across cultures 6 Given the absence of research on cultural differences in interpersonal attraction the hypotheses that we offer regarding the basis of liking for Japanese are speculative. First, to the extent that Japanese are not as concerned with elaborating a unique view of self (Kim & Markus, 1999; Markus & Kitayama, 1991), they should be less likely to seek out people who are similar to them. If the uniqueness of one’s self is not emphasized then there should be less desire to find friends who match one’s idiosyncratic qualities and can accommodate the expression of one’s uniqueness. This reasoning predicts that Japanese should not be as strongly motivated to have friends who are similar to them as Americans and they should thus show weaker relations between liking and similarity. Second, to the extent that Japanese are more context-dependent than Americans a consistent self-view would not be as desirable nor functional. The Japanese self needs to accommodate to situational pressures, and thus flexibility across situations would be more likely to be rewarded (Campbell et al., 1996; cf., Suh, 2000). To the extent that Japanese relationships and roles take precedence over abstract internal attributes, such as attitudes, traits, and abilities in self-definition (Heine, in press-b; Markus & Kitayama, 1991), having people confirm these internal qualities via self-verification would seem to be of relatively little utility. We thus predict that self-verification effects should be weaker among Japanese than Americans. Third, the mere exposure effect predicts liking based on the ease of accessibility. The more familiar we are with something, the greater ease we have recalling it from memory, and the more we like it. It would seem that this effect is largely immune from cultural effects – indeed, it has been observed across species (e.g., Zajonc, Markus, & Wilson, 1974). Japanese, like Americans, should prefer those people who are more familiar, and more easily brought to mind. We thus hypothesized no cultural differences in the mere exposure effect between cultures. Self-peer evaluations across cultures 7 Fourth, to the extent that self-enhancement is less functional for Japanese than Americans (e.g., Heine et al., 1999, in press) it would seem that Japanese would be less likely to utilize friendships as a means to self-enhance. Hence, we predict that there should be a weaker relation between liking and how positively peers evaluate the self in Japan compared to the US. Method Participants All participants belonged to a close-knit university-based organization. Our criteria for selecting groups was that the groups had to meet on a regular basis (at least once a week) and the participants had to have been active members of the group for at least one year. People were approached individually and were offered $10 each (1000 yen in Japan) if five people from their group would complete a questionnaire packet. Thus, each group received a total of $50 or 5000 yen. They were told that they should select five people from their group who were close and saw each other on a regular basis. The Japanese sample was collected at Kyoto University and the American sample at the University of Pennsylvania. We obtained the participation of a total of 10 groups of 5 people each (12 females and 38 males) from Japan and 12 groups of 5 people each (with the exception of one group that consisted of only 3 people) for a total of 58 people (32 females and 26 males) from the US. The Japanese sample consisted of 6 sports clubs, 2 international interest groups, and 2 arts clubs, whereas the US sample consisted of 4 performing arts groups, 2 sports teams, 1 fraternity, 2 sororities, 2 other organizations, and 1 group that was not associated through any particular activity. Most of the groups in both cultures were single-sex with a few exceptions. Materials and Procedure We selected 30 traits from Anderson’s (1968) list of personality traits to include in the questionnaire. We chose 10 traits that were high in desirability, 10 that were low in desirability, Self-peer evaluations across cultures 8 and 10 that were neutral, based on Anderson’s ratings. Moreover, we restricted our choice of traits to those that we felt would be comparably meaningful to both Japanese and Americans and were straightforward to translate into Japanese. We first asked students enrolled in psychology classes at Gakushuin University in Japan and at the University of Pennsylvania to evaluate each of the 30 traits in terms of their overall desirability and their importance for succeeding in society. These ratings were used in the analyses with desirability and importance reported below. We then constructed a second questionnaire based on these 30 traits. Five members of each group were given questionnaires with the names of each of the 5 members (including their own) written above columns of the 30 traits. Participants were asked to assess how well each of the traits characterized him or herself and each of his or her 4 groupmates on a scale from 1 (Extremely Inaccurate) to 15 (Extremely Accurate). Following Paulhus (1998), participants were provided with the constraint that no two individuals could receive identical scores for any of the traits. They were thus required to discriminate amongst each other. Participants were asked to indicate how much they liked each of their 4 groupmates on a scale from 1 (Dislike Intensely) to 5 (Like a lot), as a means of assessing the determinants of liking between the two cultures. We also investigated the characteristics of the groups: participants were asked to indicate how well they knew each member on a scale from 1 (Hardly at all) to 5 (Extremely Well), and how frequently they interacted with each member on a scale from 1 (Less than once every 2 weeks) to 5 (Every day). After these ratings participants were asked to complete the Rosenberg (1965) Self-Esteem Scale, Takata’s (1999) Independence/Interdependence Scale (both on 5-point Likert scales), and some demographic items. All materials were originally created in English (except Self-peer evaluations across cultures 9 for Takata’s scale which was translated from Japanese) and were translated into Japanese and checked over by two other bilinguals. Participants were instructed to complete the questionnaire in private and to never discuss their evaluations with the other members of their groups. Participants returned the completed questionnaire in a sealed envelope and were reimbursed for their participation. Results Comparability of the Samples Fifty-five percent of the American sample was female, in contrast to only 24% of the Japanese. These proportions are significantly different, 2(1, N = 108) = 10.8, p < .01. Sex was included as a factor in all analyses. The samples did not differ in terms of their average age, F(1, 106) = 2.35, ns. Americans (M = 4.00) reported knowing each other better than did Japanese (M = 2.87), F(1, 104) = 59.55, p < .001, and they also reported interacting with each other more frequently (M = 4.48) than did Japanese (M = 3.62), F(1, 105) = 28.63, p < .001. These differences are likely due to many of the American students living together – either in fraternity/sorority houses or in the same dormitory. In contrast, very few students in Japan share accommodations with other students. The two variables of “known” and “frequency” were each significantly correlated with one variable in the Japanese sample (“known” was correlated with the total positivity of self rating and “frequency” was correlated with the total positivity of rating by one’s peers; neither correlation was significant in the American sample) and are included as covariates in those analyses reported below. These two variables did not correlate with any of the other measures used in the analyses. Self-peer evaluations across cultures 10 Measures of Self-Enhancement and Self-Criticism The present data allow us to explore self-enhancement and self-criticism at the individual level. The four peer ratings, when averaged together, form a benchmark by which we can compare participants’ self-assessments. The extent to which participants’ self-evaluations depart from how they are evaluated by their peers is an index of a bias. Self-enhancement was operationalized as one’s self-evaluation being more positive than the average evaluation by one’s four peers, and self-criticism as self-evaluations being less positive than peer evaluations. To determine the positivity of the ratings, we first standardized the desirability ratings (obtained from the pretest) within each culture across the 30 traits and multiplied these standardized scores with participants’ evaluations. The 30 products of trait and standardized desirability ratings were then summed to yield a total positivity rating. One way of assessing self-enhancement is to examine the proportion of people who had positivity ratings that were greater than the average rating of them by their peers. By this standard Americans, as a group, were overall self-enhancing as 62% of them viewed themselves more positively than their peers did. This is consistent with Taylor and Brown’s (1988) contention that most “people” (Americans in this case) self-enhance. In contrast, by this standard, the Japanese sample was overall self-critical as 84% of them viewed themselves less positively than their peers did (no participants had self estimates exactly equal to their peers’ estimates). In sum, with respect to this measure of self-enhancement, the majority of Americans self-enhanced, whereas a majority of Japanese self-criticized. The proportions of self-enhancers are significantly different across the cultures, 2(1, N = 108) = 23.7, p < .001. Another test of self-enhancement was conducted via a repeated measures ANCOVA between cultures on the variables of self-positivity rating and positivity rating by peers, including the covariates “How well do you know this person?” and “How frequently do you interact with Self-peer evaluations across cultures 11 this person?” This yielded a highly significant interaction, F(1, 104) = 29.38, p < .001 (see Table 1), parallel to the above analyses with proportions. Simple effect analyses revealed that the American sample was slightly self-enhancing, in that their self-assessments were more positive than their assessments by their peers, F(1, 56) = 5.45, p < .03. In contrast, the Japanese sample was self-critical, as they viewed themselves less positively than their peers did, F(1,47) = 13.80, p < .001. Thus, this study successfully replicated the cultural differences observed in some past cultural research (e.g. Heine & Lehman, 1999; Kitayama et al., 1997; Markus & Kitayama, 1991) with an entirely different method of assessing self-enhancement and self-criticism. However the magnitude of self-enhancing tendencies among both American and Japanese samples appears to be somewhat weaker than that found in other designs. Correlates of Self-Enhancement The difference between self and peer evaluations represents one measure of self-enhancement. A more indirect index of self-enhancement can be observed in the relation between self-enhancement and trait desirability/importance. Past research has revealed that North Americans will indirectly self-enhance by evaluating themselves more positively in domains that are of special importance to them (Dunning, 1995; Heine & Lehman, 1999). To investigate this measure of self-enhancement we calculated a self-enhancement score for each trait by subtracting the evaluations of participant’s peers from their own self-evaluations. Within-participant correlations were then calculated between this self-enhancement difference score and the desirability and the importance of the traits across the 30 traits. These coefficients were converted to Fisher’s z’s before being included in analyses, and then converted back to r’s as they are reported here (McNemar, 1962). For the American sample, the average within-participant correlation between self-enhancement and desirability was r = .13, t(57) = 2.44, p < .02., and the correlation between self-enhancement and importance Self-peer evaluations across cultures 12 was r = .07, t(57) = 1.65, ns. Thus, our American sample showed evidence for this indirect means of self-enhancing by viewing themselves significantly more positively for the traits that are viewed as most desirable, and a slight, nonsignificant trend for those traits they viewed as most important. This pattern replicates that found in other studies of self-enhancement with North Americans (Dunning & Cohen, 1992; Dunning, 1995; Heine et al., in press; Heine & Lehman, 1995, 1999), although the magnitude of the correlation appears to be weaker in this design. In stark contrast, Japanese self-enhancement was correlated negatively with desirability (r = -.30, p < .001) and with importance (r = -.31, p < .001). That is, Japanese were more self-critical for those traits that were most desirable and important. The Japanese pattern of increasing self-criticism alongside desirability/ importance is consistent with that found in some other studies (e.g., Heine et al., in press; Heine & Lehman, 1999; cf., Kitayama et al., 1997), although some studies employing different methodologies with Japanese participants have found no relation (Heine & Lehman, 1995) or evidence of the opposite relation (Ito, 1999). That the present study revealed more self-enhancement among Americans for desirable traits, whereas Japanese were more self-critical for desirable and important traits, reveals a cultural difference in indirect self-enhancement as well. Within-participant correlational analyses with the two components of the self-enhancement difference score (i.e., self and peer ratings) shed light on this cultural difference in the relation between self-enhancement and trait desirability/importance. The average correlation of American self-ratings was r = .65 with desirability and r = .59 for importance (after r-z-r transformations), both ps < .001, and the average correlation of their ratings by their peers was r = .70 with desirability and r = .65 with importance, both ps < .001. That is, the more desirable and important the trait, the more positively American participants viewed themselves and were viewed by others. These correlations are likely so high as the traits Self-peer evaluations across cultures 13 covered such a broad range of desirability, and most participants were in agreement that they and their friends were better characterized by positive traits such as intelligence and considerateness than they were by negative traits such as cruel and boring. In contrast, the average correlation of Japanese self-ratings was r = .08 with desirability and r = .03 with importance, both ns, whereas the average correlation of their peer ratings was r = .54 with desirability and r = .50 with importance, both ps < .001. That is, Japanese self-assessments were unaffected by the desirability and importance of the trait (i.e., they were just as likely to say that they were intelligent and considerate as they were to say that they were cruel and boring), however, as in the American case, their peers tended to view them more positively for traits high in desirability and importance. We also used the self-enhancement difference scores that we had calculated for the above analyses to investigate the relations between positive views of the self and cultural values. We correlated individual’s total self-enhancement difference score with the three trait measures that we had collected (viz., self-esteem, independence, and interdependence; see Table 2). Overall, positive self-views as operationalized by both self-esteem and self-enhancement scores appear to be associated with independence and interdependence in ways consistent with past findings (e.g., Heine et al., 1999; Kiuchi, 1996; Yamaguchi, 1994). That is, both Americans and Japanese who scored higher in independence had higher self-esteem and tended to view themselves more positively than did their peers. Likewise, those who scored higher on interdependence showed lower self-esteem and did not tend to show self-enhancement biases (this latter relation was evident only in the American sample). These relations within cultures thus parallel the differences between cultures. Thus, self-enhancement can be unpackaged, at least in part, in terms of independence and interdependence. Self-peer evaluations across cultures 14 Accuracy Analyses There are multiple ways to operationalize interpersonal accuracy. One approach is to assess the absolute value of the discrepancy between self- and peer-estimates with smaller discrepancies indicating greater accuracy. However, the magnitude of discrepancy scores is confounded by “elevation” (Cronbach, 1955). For example, an individual rater who tends to give higher evaluations of others will create self-peer discrepancy scores that are of a different magnitude than those of a rater who tends to give lower evaluations, thus compromising this as an index of accuracy (Kenny, 1993). A second approach is to calculate the correlation between self- and peer-ratings with larger correlation coefficients indicating greater accuracy. Relations between self- and peer-ratings can be calculated in a number of ways. One method is to calculate intraclass correlations between self- and peer-ratings (e.g., Shrout & Fleiss, 1979). Although this method is often favored in accuracy analyses, it cannot be used when raters are assessing different targets (Kenny, 1993), as they are here. A second method is to use a round robin analysis of variance (e.g., Warner, Kenny, & Stoto, 1979). However, because we are not measuring interactions between peers and self, but are just comparing self and other-evaluations, it is not necessary to separate the interaction terms from the ANOVA model. A third method, which has been used in similar studies (e.g., John & Robins, 1993) is to calculate within-participant correlations between self- and peer-ratings. We utilized this operationalization of accuracy in the present study. A within-participant correlation coefficient was calculated for each participant between their self-rating and the average rating by their four peers across all 30 traits. The average within-participant self-peer correlation (after r-z-r transformations) for Americans was r = .70, whereas the average correlation for Japanese was r = .32, both ps < .001. An ANOVA revealed that the American correlation was significantly higher than that of the Japanese, F(1, 104) = Self-peer evaluations across cultures 15 60.50, p < .001. That is, American self-assessments were more similar to those of their peers across the 30 traits than were Japanese. We also analyzed a second index of accuracy: the absolute value of the discrepancy scores between self evaluations and evaluations by peers. An ANOVA conducted on these difference scores likewise reveals that Americans were more accurate than Japanese: the American average difference score (M = 2.60; SD = .61) was smaller than that of Japanese (M = 3.26; SD = 1.13), F(1,106) = 14.68, p < .001, thereby replicating the correlational analyses. It thus does not appear to be the case that Japanese self-enhance less than Americans because they are more attuned to how others view them. On the contrary, Americans self-assessments were more in agreement with those of their peers than were Japanese. We assessed whether the desirability of the trait was related to the accuracy of assessment. We first conducted median splits on the desirability of the 30 traits, and conducted analyses separately for the 15 negative and the 15 positive traits. We calculated the average self-peer correlation coefficient for each trait, and then correlated these with the average desirability scores for each trait. For these analyses the unit of analysis was the trait. Figure 1 reveals that for Americans, self-peer correlations were strongly tied to the desirability of the traits replicating John and Robins (1993). For negative traits, Americans showed a positive correlation between accuracy and the desirability of the traits, r = .56, p < .08. In contrast, for positive traits Americans showed a negative correlation between accuracy and desirability, r = -.64, p < .05. Examining the data in another light, we created a measure of the overall evaluativeness of the traits by taking the absolute value of the standardized desirability scores. Across all 30 traits the correlation between evaluativeness and accuracy was r = -.56, p < .001. In sum, there is much agreement between American self and peer assessments for traits of neutral valence, Self-peer evaluations across cultures 16 however there is little agreement for traits that are either high or low in desirability. A few examples illustrate this: Americans rated “sincerity” as highly desirable (8.3 out of 9) and the correlation between self and peer ratings was r = .00. “Selfish” was rated as highly undesirable (2.1 out of 9) and the self-peer correlation was r = -.03. In contrast, “impulsive” was rated as neutral (4.9 out of 9), and its correlation between self and peer assessments was r = .49. In sum, there is much self-peer agreement for traits are evaluatively neutral. When strongly desirable or undesirable traits are considered, in contrast, there is strikingly little agreement. This is evidence that self-protective motivations are clouding American individuals’ accuracy in evaluating themselves. These findings closely replicated those of John and Robins (1993). In contrast, Japanese showed no relation between accuracy and desirability for either negative (r = -.02, ns), or positive traits, (r =-.22, ns; See Figure 2). Evaluativeness and accuracy showed a slight negative correlation, however it was not significant, r = -.14, ns. That is, the valence of the trait did not bear on self-peer agreement for Japanese. Self-protective motivations do not appear to influence the accuracy in Japanese self-assessments. The correlations between evaluativeness and accuracy were marginally different between cultures, p < .08. Correlates of Liking We calculated correlations between how much individuals liked their peers and a number of liking-relevant variables to test which theories of interpersonal attraction are supported within the two cultures. For these analyses we used the individual’s evaluation of the relationship as the unit of the analysis (i.e., each person rated 4 peers so there are 4n units of liking) and all correlations were first converted to Fisher’s z’s. First, we explored how well the similarity effect predicted liking. We had two measures of similarity, perceived similarity (how similar the target perceives their relationship partner to be) and actual similarity (how similar the two Self-peer evaluations across cultures 17 relationship partners’ self-rated personalities actually are). An examination of perceived similarity (i.e., the correlation between liking and the average within-participant correlation between the participant’s self-assessment and the participant’s evaluation of his or her peer) revealed that liking and perceived similarity exhibited a significant positive correlation in the American sample (see Table 3). Americans preferred friends who were perceived as similar to themselves replicating past research (e.g., Byrne, 1971). In contrast, these two variables were not related in the Japanese sample. Japanese did not show a significant preference for friends who were viewed as similar to them (at least in terms of personality – similarity of other variables such as attitudes or interests may be predictive of liking for Japanese). These correlations are significantly different across cultures, p < .001. In contrast, an examination of actual similarity (i.e., the correlation between liking and the average within-participant correlation between each partner’s self-evaluation) did not reveal a significant relation for either Americans or Japanese. The two correlations were not significantly different (p > .20) although the nominal trend of the correlations paralleled the cultural difference in perceived similarity Second, we explored how much liking of a peer correlated with self-verification (i.e., the average within-participant correlation between the participant’s self-rating and the rating of the participant by his or her peer). A significant positive correlation emerged between liking and self-verification for the American sample. Americans prefer friends who see them in the same ways that they see themselves, replicating past research (e.g., Swann et al., 1992). Japanese, in contrast, showed no correlation between these two variables. The extent to which others share one’s own view of oneself apparently has little bearing on liking within close groups in Japan. This absence of a relation further supports the notion that consistent self-views are less functional in East Asian cultures compared with North Americans (Campbell et al., 1996; Self-peer evaluations across cultures 18 Kanagawa, Markus, & Cross, 2001; Suh, 2000). These two correlations differed across cultures, p < .001. Third, we explored the correlations between liking and the mere exposure effect: that is, how much participants liked those whom they knew well and interacted with frequently.. Although the mere exposure effect tends to be a curvilinear phenomenon (Berlyne, 1967), and our measures of the degree of exposure are not particularly rigorous (responses were made on 5-point subjective Likert scales), both cultures showed clear evidence of this effect replicating past research (e.g., Zajonc, 1968). The effect was stronger for the question “How well do you know this person?” than it was for the question “How frequently do you interact with this person?”, although both correlations were significant within each culture. Interestingly, both correlations were identical for the two cultures to the second decimal place. Fourth, we explored the relations between liking and reflected self-enhancement: that is, how much participants liked those who viewed them positively (i.e., the peer’s overall positivity rating of the participant). The American sample showed a significant positive correlation between liking and reflected self-enhancement. Americans tended to prefer those people who viewed them positively. This replicates past research (e.g., Murray et al., 1996). Japanese also showed a significant positive correlation, although this relation was smaller than it was for the American sample, p < .05. Thus, Japanese also prefer those who view them positively. That the relation was smaller for the Japanese further suggests that self-enhancement is less functional for Japanese compared with Americans. We also explored relations between how much participants were liked overall by their peers and independence/ interdependence. Interestingly, independence was negatively correlated with being liked, and interdependence was positively correlated with being liked for Japanese, Self-peer evaluations across cultures 19 but not for Americans. There may be stronger interpersonal consequences in Japan for highlighting one’s separation or connection with others. In sum, Americans liked those who they perceived to be similar to them, who shared their own view of themselves, who they interacted with frequently, and who viewed them positively. Americans’ liking scores were unrelated to their actual similarity, independence, or interdependence. For Japanese, liking was positively associated with familiarity, interdependence, and reflected self-enhancement (although the relation with the latter was weaker than it was for Americans), was negatively associated with independence, and was unrelated with perceived and actual similarity, and self-verification. These findings, although exploratory, suggest some different bases in liking across cultures. Discussion The present results replicate the cultural differences in self-enhancement observed in many past studies and challenge the notion that the weaker self-enhancing tendencies observed among Japanese are due to them being more accurate in their self-perceptions than Americans. Rather, in this study, American self-assessments were significantly more aligned with how their peers viewed them than were those of Japanese. This cultural difference in accuracy was highly pronounced and warrants further investigation. A recent study by Suh (2000) similarly found lower agreement between self- and other-assessments for Koreans than among Americans. Suh speculated that this cultural difference in interjudge agreement might be due to cultural differences in consistency of behaviors across situations. To the extent that individuals are revealing different aspects of themselves to different people, the individual’s self-view will contain information regarding the self that the peers might never have seen, and thus agreement should be lower (Colvin, 1993). In other words, the lower self-peer agreement in our Japanese sample may have arisen from Japanese individuals rating others based on their “school club Self-peer evaluations across cultures 20 self,” whereas they were evaluating their own self based on how they have felt and acted across all contexts. To the extent that Americans are more consistent than Japanese across situations (e.g., Heine, in press-b; Kanagawa et al., 2001; Suh, 2000), there should be less discrepancy in the information that peers and selves have, thus enhancing self-peer agreement. That Japanese were not more accurate in their self-perceptions provides further support that the relative lack of self-enhancement observed among Japanese in past studies is due to self-critical motivations. Indeed, there was much evidence of self-criticism among the Japanese participants in the present study. The Japanese sample tended to view themselves more negatively than their peers did, and this self-criticism was stronger for the traits viewed as most desirable and important, in stark contrast to the pattern exhibited within the American sample. Moreover, the correlations between self and peer ratings exhibited by Japanese lacked the self-protective motivations evident in the American self-assessments. Americans and their peers only agreed on their ratings for neutral traits, whereas they showed almost no agreement on evaluatively laden traits. Japanese accuracy, in contrast, was unaffected by the valence of the traits. Most past research on cultural differences in self-enhancement has only explored effects at the group level and has thus obscured an analysis of the dynamics at the individual level. Although Japanese as a group often appear to be self-critical (Heine et al., 1999), it is possible that these group scores derive from a small minority of highly self-critical Japanese among a majority of mild self-enhancers. The present study allowed us to determine which individuals were being self-critical, and revealed that the vast majority of Japanese viewed themselves more negatively than did their peers. Self-criticism among Japanese individuals is a widely shared tendency, at least among Kyoto University students. Likewise, a majority of our American Self-peer evaluations across cultures 21 sample was self-enhancing. These divergent self-assessment styles thus appear to be widely shared among members of the two cultures. Being able to identify who is self-enhancing is a necessary step for investigating the benefits or detriments of self-enhancement (Colvin et al., 1995; Paulhus, 1998). In the present study, self-enhancement and self-esteem tended to correlate with independence and interdependence within cultures in ways that paralleled the differences between the cultures. Those individuals who embraced more independent views of the self relative to their compatriots were more self-enhancing, whereas those with more interdependent views were less self-enhancing, for both Americans and Japanese. This replicates past findings (e.g., Kiuchi, 1996; Yamaguchi, 1994 ) and increases our confidence that self-enhancement is associated with autonomy and independence (Heine et al., 1999). The present data also investigated cross-cultural bases of liking. Interestingly, there were some striking divergences on what leads to liking in the two cultures. These data on liking are novel, and the dependent measures were limited to subjective Likert scale responses, and thus these analyses should be viewed as preliminary and the conclusions speculative until they are replicated. The data speak more to bases of liking that Japanese do not appear to have (viz., perceived similarity and self-verification effects) than they do to the bases that characterize Japanese interpersonal attraction. There are likely other variables that are more predictive of liking for Japanese than Americans that were not investigated here. Nonetheless, the present pattern is suggestive of important cultural differences in relationships and future research in this area would likely prove fruitful. Limitations and Future Directions We operationalized self-enhancement and self-criticism as the extent to which self-evaluations departed from peer evaluations. This operationalization carries the assumption Self-peer evaluations across cultures 22 that the peer evaluations are a reliable and objective standard. Clearly, this assumption may not be accurate. We feel that a strength of this research is the addition of a benchmark to compare individuals’ self-evaluations, however, there may be better standards that could be used. Future research is necessary to determine whether comparisons with other benchmarks yield parallel findings to those obtained here. All of the findings here are based on self-report measures. It is thus possible that the obtained effects are colored by self-presentation motives, with American self-enhancement being the product of false bravado and Japanese self-criticism emerging from false modesty. We cannot rule this possibility out with the present dataset. However, that other research finds comparable cultural differences in studies employing hidden measures render this account less plausible (e.g., Heine et al., in press; Heine & Lehman, 1997; Heine, Takata, & Lehman, 2000; Meijer, 2000). Nonetheless, the present results would be more compelling if replicated with hidden behavioral measures (e.g., the design of Heine et al, 2000 could be used in such a study). Many of the methodologies employed in social and personality psychology stem from the view that individual selves are complete in and of themselves and thus can best be studied in isolation from any contaminating social influences. However, to the extent that interdependent selves are sustained by roles and relationships (Heine, in press-b; Markus & Kitayama, 1991), investigating the individual self in isolation or in the context of interactions with strangers will only reveal a misleading fraction of the self (Lebra, 1976). The interdependent self is clearly not interdependent with everyone; rather, significant relationships represent a particular class of ingroup relations (e.g., Heine, in press-b; Iyengar, Lepper, & Ross, 1999). Indeed, some research reveals that whereas East Asians are particularly close with ingroup members they are even more distant from outgroup members than Americans (e.g., Iyengar et al., 1999; Yamagishi & Yamagishi, 1994). We feel that investigations of the interdependent self will provide the Self-peer evaluations across cultures 23 most accurate analysis if the self is studied in the context of the naturally-occurring and significant relationships that sustain it. Self-peer evaluations across cultures 24 References Anderson, N. H. (1968). Likableness ratings of 555 personality-trait words. Journal of Social Psychology, 9, 272-279. Berlyne, D. E. (1967). Arousal and reinforcement. In D. Levine (Ed.), Nebraska symposium on motivation, (pp. 1-110). Lincoln: University of Nebraska Press. Byrne, D. (1971). The attraction paradigm. New York: Academic Press. Campbell, J. D., Trapnell, P., Heine, S. J., Katz, I. M., Lavallee, L. F., & Lehman, D. R. (1996). Self-concept clarity: Measurement, personality correlates, and cultural boundaries. Journal of Personality and Social Psychology, 70, 141-156. Chang, H.-C., & Holt, G. R. (1994). A Chinese perspective on face as inter-relational concern. In S. Ting-Toomey (Ed.), The challenge of facework: Cross-cultural and interpersonal issues (pp. 95-132). Albany, NY: SUNY Press. Cohen, D., & Gunz, A. (2001). As seen by the other… The self from the “outside in” and the “inside out” in the memories and emotional perceptions of Easterners and Westerners. Manuscript submitted for publication. University of Waterloo. Colvin, C. R. (1993). Judgable people: Personality, behavior, and competing explanations. Journal of Personality and Social Psychology, 64, 861-873. Colvin, C. R., Block, J., & Funder, D. C. (1995). Overly positive evaluations and personality: Negative implications for mental health. Journal of Personality and Social Psychology, 68, 1152-1162. Cronbach, L. J. (1955). Processes affecting scores on “understanding of others” and “assumed similiarity.” Psychological Bulletin, 52, 177-193. Dion, K. K., Dion, K. L. (1993). Individualistic and collectivistic perspectives on gender and the cultural context of love and intimacy. Journal of Social Issues, 49 (3), 53-69. Self-peer evaluations across cultures 25 Dunning, D. (1995). Trait importance and modifiability as factors influencing self-assessment and self-enhancement motives. Personality and Social Psychology Bulletin, 21, 1297-1306. Dunning, D., & Cohen, G. L. (1992). Egocentric definitions of traits and abilities in social judgment. Journal of Personality and Social Psychology, 63, 341-355. Harrison, A. A. (1977). Mere exposure. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology, (pp. 39-83). New York: Academic Press. Heine, S. J. (in press-a). An exploration of cultural variation in self-enhancing and self-improving motivations. Nebraska Symposium of Motivation. Heine, S. J. (in press-b). Self as cultural product: An examination of East Asian and North American selves. Journal of Personality. Heine, S. J., Kitayama, S., Lehman, D. R., Takata, T., Ide, E., Leung, C., & Matsumoto, H. (in press). Divergent consequences of success and failure in Japan and North America: An investigation of self-improving motivations and malleable selves. Journal of Personality and Social Psychology. Heine, S. J. & Lehman, D. R. (1995). Cultural variation in unrealistic optimism: Does the West feel more invulnerable than the East? Journal of Personality and Social Psychology, 68, 595-607. Heine, S. J., & Lehman, D. R. (1997). Culture, dissonance, and self-affirmation. Personality and Social Psychology Bulletin, 23, 389-400. Heine, S. J., & Lehman, D. R. (1999). Culture, self-discrepancies, and self-satisfaction. Personality and Social Psychology Bulletin, 25, 915-925. Heine, S. J., Lehman, D. R., Markus, H. R., & Kitayama, S. (1999). Is there a universal need for positive self-regard? Psychological Review, 106, 766-794. Self-peer evaluations across cultures 26 Heine, S. J., Takata, T., & Lehman, D. R. (2000b). Beyond self-presentation: Evidence for self-criticism among Japanese. Personality and Social Psychology Bulletin, 26, 71-78. Holloway, S. D. (1988). Concepts of ability and effort in Japan and the United States. Review of Educational Research, 58, 327-345. Ito, T. (1999). Self-enhancement tendency and other evaluations: An examination of "better-than-average effect.' (In Japanese). Japanese Journal of Psychology, 70, 367-374. Iyengar, S. S., Lepper, M. R., & Ross, L. (1999). Independence from whom? Interdependence with whom? Cultural perspectives on ingroups versus outgroups. In D. Miller & D. Prentice (Eds.), Cultural Divides: Understanding and overcoming group conflict (pp. 273-301). New York: Sage. John, O. P., & Robins, R. W. (1993). Determinants of interjudge agreement on personality traits: The big five domains, observability, evaluativeness, and the unique perspective of the self. Journal of Personality, 61, 521-551. John, O. P., Robins, R. W. (1994). Accuracy and bias in self-perception: Individual differences in self-enhancement and the role of narcissism. Journal of Personality & Social Psychology, 66, 206-219. Kanagawa, C., Cross, S. E., & Markus, H. R. (2001). “Who am I?”: The cultural psychology of the conceptual self. Personality and Social Psychology Bulletin, 27, 90-103. Kenny, D. A. (1993). A coming-of-age for research on interpersonal perception. Journal of Personality, 64, 789-807. Kim, H., & Markus, H. R. (1999). Deviance or uniqueness, harmony or conformity? A cultural analysis. Journal of Personality and Social Psychology, 77, 785-800. Self-peer evaluations across cultures 27 Kitayama, S., Takagi, H., & Matsumoto, H. (1995). Causal attribution of success and failure: Cultural psychology of the Japanese self. Japanese Psychological Review, 38, 247-280. (In Japanese). Kiuchi, A. (1996). Independent and interdependent construals of the self: Cultural influences and relations to personality traits. Japanese Journal of Psychology, 67, 308-313. (In Japanese). Lebra, T. S. (1976). Japanese patterns of behavior. Honolulu: University of Hawaii Press. Markus, H. R., & Kitayama, S. (1991). Cultural variation in the self-concept. In G. R. Goethals & J. Strauss (Eds.), Multidisciplinary perspectives on the self. (pp. 18-48). New York: Springer-Verlag. McNemar, Q. (1962). Psychological statistics (3rd ed.). New York: Wiley. Meijer, Z. Y. (2000). Automatic selves across cultures: A comparative study in Japan and the United States. Unpublished doctoral dissertation. University of Virginia. Morling, B., Kitayama, S., & Miyamoto, Y. (in press). Control and adjustment as two modes of constructing social situations: A comparison between Japan and the United States. Personality and Social Psychology Bulletin. Murray, S. L., Holmes, J. G., & Griffin, D. W. (1996). The benefits of positive illusions: Idealization and the construction of satisfaction in close relationships. Journal of Personality and Social Psychology, 70, 79-98. Paulhus, D. L. (1998). Interpersonal vs. intrapsychic adaptiveness of trait self-enhancement: A mixed blessing? Journal of Personality and Social Psychology, 74, 1197-1208. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton: Princeton University Press. Self-peer evaluations across cultures 28 Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420-428. Suh, E. M. (2000). Culture, identity consistency, and subjective well-being. Manuscript submitted for publication. University of California, Irvine. Swann, W. B., Jr., Stein-Seroussi, A., & Giesler, R. B. (1992). Why people self-verify. Journal of Personality and Social Psychology, 62, 392-401. Takata, T. (1999). Developmental process of independent and interdependent self-construal in Japanese culture: Cross-cultural and cross-sectional analyses (In Japanese). Japanese Journal of Educational Psychology, 47, 480-489. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193-210. Warner, R. M., Kenny, D. A., & Stoto, M. (1979). A new round robin analysis of variance for social interaction data. Journal of Personality and Social Psychology, 37, 1742-1757. Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39, 806-820. Yamagishi, T., & Yamagishi, M. (1994) Trust and commitment in the United States and Japan. Motivation and Emotion, 18, 9-66. Yamaguchi, S. (1994). Collectivism among the Japanese: A perspective from the self. In U. Kim, H. C. Triandis, C. Kagitcibasi, S. -C. Choi & G. Yoon (Eds.) Individualism and Collectivism: Theory, method, and applications. (pp. 175-188). Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology (Monograph Suppl., Pt. 2), 1-29. Zajonc, R. B., Markus, H., & Wilson, W. R. (1974). Exposure, object preference, and distress in the domestic chick. Journal of Comparative and Physiological Psychology, 86, 581-585. Self-peer evaluations across cultures 29 Author’s Notes This research was funded by start-up funds from the University of Pennsylvania and a fellowship from the Japan Foundation to the first author. We thank Michiko Sakuma for her help in getting pretest data from Gakushuin University, and Mamiko Hada and Nariko Takayanagi for their help with the translations. Correspondence regarding this article should be addressed to Steven J. Heine at the Department of Psychology, 2136 West Mall, University of British Columbia, Vancouver, BC, V6T 1Z4 Canada. E-mail can be sent to [email protected]. Self-peer evaluations across cultures 30 Table 1. Means and standard deviations of key dependent measures Americans Japanese Positivity of 2.49a .17b Self-Assessments (1.01) (1.35) Positivity of 2.03a 1.27b Assessments by Peer (1.10) (.94) .46a -1.10b (1.38) (1.57) .70a .32b (.16) (.29) 4.17a 3.32b (.58) (.90) 3.68a 2.93b (.51) (.54) 3.37a 3.74b (.61) (.54) Difference Self-Peer Correlations Self-Esteem Independence Interdependence Columns with different subscripts are significantly different at p < .01. Self-peer evaluations across cultures 31 Table 2. Correlations between self-enhancement and trait measures Self-Enhancement Self-Esteem Independence Interdependence .15 .31* -.25* .52*** -.44*** Difference Scores Self-Enhancement Difference Scores Self-Esteem --- .46*** --- Independence .30* .51*** -.35** --- Interdependence -.02 -.38** -.39** --- Correlations above the diagonal are for Americans, and those below for Japanese * p < .06 ** p < .01 *** p < .001 Self-peer evaluations across cultures 32 Table 3 Correlates of Liking Americans Japanese .55a*** .10b .11a -.02a .33a*** -.00b .37a*** .37a*** .17a* .17a* .40a*** .22b** Independence -.08a -.34a* Interdependence .19a .32a* Similarity Effect I (Perceived Similarity)+ (Average within-participant correlation between participants’ ratings of self and peer) Similarity Effect II (Actual Similarity)+ (Average within-participant correlation between actual ratings of self by both partners) Self-Verification+ (Average within-participant correlation between self rating and rating by peer) Mere Exposure Effect I+ (How well do you know this person?) Mere Exposure Effect II+ (How frequently do you interact with this person?) Reflected Self-Enhancement+ (Degree of positivity of rating by peer) Note: + These analyses are conducted with the individual relationship as the unit of analysis. Columns with different subscripts are significantly different at p < .05. * p < .05 ** p < .01 *** p < .001 Self-peer evaluations across cultures 33 Figure Captions 1. Self-Peer Agreement versus Desirability for the American sample. 2. Self-Peer Agreement versus Desirability for the Japanese sample. Self-peer evaluations across cultures 34 Self-peer evaluations across cultures 35 Self-peer evaluations across cultures 36 Self-peer evaluations across cultures 37